metadata

title: gradio-experiments
emoji: 🔬
colorFrom: gray
colorTo: blue
sdk: gradio
sdk_version: 5.19.0
suggested_hardware: cpu-basic
app_file: src/app.py
pinned: true
disable_embedding: true
license: mit

gradio-experiments

A collection of feature experiments with Gradio.

Environment variables

You can specify the following environment variables in a .env, for example. In addition to those, Gradio environment variables can also be specified but GRADIO_SERVER_NAME (set to "0.0.0.0"), GRADIO_SHARE (set to False) and GRADIO_ANALYTICS_ENABLED (set to False) will be ignored.

Variable	Default value	Description
LOCAL_STORAGE_ENCRYPTION_KEY	None	If you want to be able read from previously saved local storage on the browser, you must specify this environment variable. If it is not specified, a new value is generated for a browser session.

Installation and use

Create a Python virtual environment for Python 3.12.0 or above. In that virtual environment, run pip install -U -r requirements.txt. (You can clean everything you have installed in that environment by running python -m pip freeze | cut -d "@" -f1 | xargs pip uninstall -y.) Once all the packages are installed, you can start the Gradio server by running python src/app.py. If the server is able to start without any problem, it will be listening to port 7860 on any IP address of your network adapter. You will be able to see the web interface on your localhost at http://127.0.0.1:7860/.

A public deployment may be available through Huggingface Spaces.

Status

The following features have been or are being implemented.

Feature	Description	Status
State management	Demonstration of global, session and browser (local) states.	Completed
Datasets	Experimentation with datasets and data frames.	Preliminary stage
Pydantic enitity profile	Experimentation with Pydantic based entity profiles.	Partially implemented
Text transformation	Just a placeholder.	Completed